-
RESEARCH Open Access
Health-related quality of life in relation tosymptomatic and
radiographic definitionsof knee osteoarthritis: data
fromOsteoarthritis Initiative (OAI) 4-year follow-up studySoili
Törmälehto1,2* , Mika E. Mononen2, Emma Aarnio1,3, Jari P. A.
Arokoski4,5, Rami K. Korhonen2,6
and Janne Martikainen1
Abstract
Background: The purpose was to quantify the decrement in health
utility (referred as disutility) associated with kneeosteoarthritis
(OA) and different symptomatic and radiographic uni- and bilateral
definitions of knee OA in a repeatedmeasures design of persons with
knee OA or at increased risk of developing knee OA.
Methods: Data were obtained from the Osteoarthritis Initiative
database. SF-12 health-related quality of life was convertedinto
SF-6D utilities, and were then handled as the health utility loss
by subtracting 1.000 from the utility score, yielding anegative
value (disutility). Symptomatic OA was defined by radiographic
findings (Kellgren-Lawrence, K-L, grade≥ 2) andfrequent knee pain
in the same knee. Radiographic OA was defined by five different
definitions (K-L≥ 2 unilaterally /bilaterally, or the highest /
mean / combination of K-L grades of both knees). Repeated measures
generalized estimatingequation (GEE) models were used to
investigate disutility in relation to these different
definitions.
Results: Utility decreased with worsening of symptomatic or
radiographic status of knee OA. The participants withbilateral and
unilateral symptomatic knee OA had 0.03 (p < 0.001) and 0.02 (p
< 0.001) points lower utility scores,respectively, compared with
the reference group. The radiographic K-L grade 4 defined as the
mean or the highestgrade of both knees was related to a decrease of
0.04 (p < 0.001) and 0.03 (p < 0.001) points in utility
scores,respectively, compared to the reference group.
Conclusions: Knee OA is associated with diminished
health-related quality of life. Health utility can be quantified
inrelation to both symptomatic and radiographic uni- and bilateral
definitions of knee OA, and these definitions areassociated with
differing disutilities. The performance of symptomatic definition
was better, indicating that painexperience is an important factor
in knee OA related quality of life.
Keywords: Knee osteoarthritis, Health-related quality of life,
SF-12, SF-6D
* Correspondence: [email protected] and
Outcomes Research Unit (PHORU), School ofPharmacy, University of
Eastern Finland, Kuopio, Finland2Department of Applied Physics,
University of Eastern Finland, Kuopio,FinlandFull list of author
information is available at the end of the article
© The Author(s). 2018 Open Access This article is distributed
under the terms of the Creative Commons Attribution
4.0International License
(http://creativecommons.org/licenses/by/4.0/), which permits
unrestricted use, distribution, andreproduction in any medium,
provided you give appropriate credit to the original author(s) and
the source, provide a link tothe Creative Commons license, and
indicate if changes were made. The Creative Commons Public Domain
Dedication
waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies
to the data made available in this article, unless otherwise
stated.
Törmälehto et al. Health and Quality of Life Outcomes (2018)
16:154 https://doi.org/10.1186/s12955-018-0979-7
http://crossmark.crossref.org/dialog/?doi=10.1186/s12955-018-0979-7&domain=pdfhttp://orcid.org/0000-0003-0138-2285mailto:[email protected]://creativecommons.org/licenses/by/4.0/http://creativecommons.org/publicdomain/zero/1.0/
-
BackgroundOsteoarthritis (OA) is a chronic and degenerative
jointdisease. It is the most common type of arthritis, and
itaffects most frequently hips, knees, and hands [1]. KneeOA is a
major cause of disability and loss of functionamong older adults,
and it causes a major burden bothto individuals and health care
systems [2, 3]. Older ageand obesity are significant risk factors
of knee OA [4, 5],and with the continuously aging population and
increas-ing prevalence of obesity, the burden of knee OA is
an-ticipated to increase.While knee OA manifests pain, stiffness,
and daily ac-
tivity deficits, it causes deterioration in
patient-reportedhealth-related quality of life (HRQoL). OA is
associatedwith strong negative effect on HRQoL [6, 7], and
bilat-eral knee pain, other joint pain comorbidity, and inad-equate
pain relief in conjunction with knee OA havebeen shown to be
associated with even poorer quality oflife [8–10]. Correspondingly,
total knee replacement hasbeen reported to improve patients’
quality of life [11].In chronic conditions such as knee OA, HRQoL
is one
of the most commonly used patient-reported outcomemetrics. In
knee OA research literature, several diseasespecific instruments
have been used under the label ofHRQoL. The Western Ontario and
McMaster Univer-sities Osteoarthritis Index (WOMAC) [12] and
TheKnee injury and Osteoarthritis Outcome Score (KOOS)[13], and
their subscales, are widely used questionnairesto assess
patient-reported outcomes related to knee OA inclinical trials.
WOMAC questionnaire assesses pain, stiff-ness, and functional
limitation [12], while KOOS includesseparate subscales for pain,
symptoms, function in dailyliving, function in sport and
recreation, and knee-relatedquality of life [13]. The benefit of
these disease-spesific in-struments is that they cover dimensions
relevant to kneeOA [14]. However, disease-spesific HRQoL measures
donot provide preference-based utility values needed inhealth
economic analyses. Instead, single, general, and cal-ibrated
preference-based health utility scores are neededto incorporate the
quantity of life (years) and quality of lifeinto quality-adjusted
life years (QALYs) [15]. QALY iscommonly used as an outcome measure
in health eco-nomic analyses and health care decision-making.In
health economic analyses, information on how health
utility values response to the disease-spesific health statesis
essential. However, in the case of knee OA, the goldstandard of
knee OA definition is currently unavailable,and variety of criteria
and definitions have been used inprevious research studies [16].
Previously, EQ-5D (Euro-Qol-5 dimensions questionnaire) health
utility values inrelation to different knee OA definitions have
beenassessed in two separate studies [17, 18]. Both of the stud-ies
used an unilateral knee OA definition. Consequently,the information
on health utility in relation to different
definitions of knee OA is scarce, and particularly informa-tion
on health utility in relation to bilateral definition ofknee OA is
lacking. Therefore, the objective of the presentstudy was to
quantify the preference-based health utilityvalues associated with
different symptomatic and radio-graphic uni- and bilateral
definitions of knee OA in a re-peated measures design of persons
with symptomaticknee OA or at increased risk of developing knee
OA.
MethodsData sourcesData used in the preparation of this article
were obtainedfrom the open access Osteoarthritis Initiative (OAI)
study(http://www.oai.ucsf.edu/), which is a multi-center,
longitu-dinal cohort study on knee OA [19]. The study incorporatesa
progression, an incidence, and a reference sub-cohort. Sub-jects in
the progression sub-cohort have symptomatic kneeOA at baseline,
while subjects in the incidence sub-cohortare at increased risk of
developing it. Reference sub-cohortsubjects have neither
symptomatic knee OA nor eligibilityrisk factors at baseline. The
specific eligibility risk factors andethical issues are described
in detail in Osteoarthritis Initia-tive Study Protocol [19].
Ethical approval for collecting allsubject information was provided
by the OAI. Informed con-sent was obtained from all individual
participants included inthe study. Applied OAI datasets are listed
in additional file(see Additional file 1). Data from baseline and
follow-upvisits at 12, 24, 36 and 48 months were applied in the
study.
Preference-based health utility indexA mapping algorithm
(https://www.sheffield.ac.uk/) wasused to convert 12-item Short
Form Health Survey (SF-12)data from the OAI study into SF-6D
utility scores [20].SF-12 is a measure of general health covering
eight healthdomains. The SF-6D estimates a preference-based
singleindex utility measure from SF-12 using general
populationvalues and standard gamble valuation technique.
SF-6Dscores fall on the scale from 0 (death) to 1 (perfect
health).The worst SF-6D score, excluding death, is 0.291 (‘floor
ef-fect’). We handled SF-6D utilities as decrement in healthutility
by subtracting 1.000 from the SF-6D utility score,yielding a
negative value (referred as disutility).
Definitions of knee OAKellgren-Lawrence (K-L) radiographic
system classifiesknee OA into five grades based on the severity of
radio-graphic findings of joint space narrowing,
osteophytes,sclerosis, and bone deformity [21]. K-L grade 0
indicatesintact joint without any features of OA, and K-L grade 2is
considered the cut-off point of definite OA, while K-Lgrade 4
indicates severe OA. However, persons with K-Lgrade ≥ 2 may be
asymptomatic and vice versa. There-fore, symptomatic definition of
knee OA is consideredmore relevant. In prevalence studies,
symptomatic knee
Törmälehto et al. Health and Quality of Life Outcomes (2018)
16:154 Page 2 of 12
http://www.oai.ucsf.edu/https://www.sheffield.ac.uk
-
OA is usually defined as the concurrent presence ofradiographic
findings (usually K-L grade ≥ 2) and fre-quent knee pain in the
same knee [1, 22, 23].In order to estimate the effect of different
definitions on
the HRQoL associated with knee OA, seven definitions ofknee OA
were studied. Two definitions were based on thesymptomatic
definition of knee OA (K-L grade ≥ 2 andknee pain, aching or
stiffness on more than half the daysduring past month in the same
knee): (a) 2-scale: no/yes(OA present either uni- or bilaterally),
or (b) 3-scale: no/unilateral/bilateral knee OA. The remaining five
defini-tions were based solely on the radiographic definition
ofknee OA: (c) 2-scale: no/yes (K-L grade ≥ 2 either uni-
orbilaterally), (d) 3-scale: no/unilateral/bilateral K-L grade ≥2,
(e) 5-scale unilateral definition: the highest K-L grade ofboth
knees, (f) 9-scale bilateral definition: mean K-L gradeof both
knees, or (g) 15-scale bilateral definition: combin-ation of K-L
grades of both knees ((0;0), (1;0), (1;1) …(4;2), (4;3), (4;4)).
Missing data regarding K-L grades wasnot imputed.
CovariatesWe used 11 covariates as adjusting variables.
Demo-graphics (age, gender, education, living status),
clinicalstatus (injuries, surgical history, body mass index
(BMI),comorbid conditions, smoking status), and physical ac-tivity
were selected as a standard set of adjusting vari-ables according
to Rolfson and co-authors [24]. We alsoincluded race into adjusting
variables because of its asso-ciation with pain level and health
status [25, 26].Injury status and surgical history were both
dichoto-
mized and based on the OAI query on ever (baseline) orsince last
visit (follow-up) injuring either knee badlyenough to limit ability
to walk for at least 2 days, and everhaving knee surgery or
arthroscopy (answers regarding leftand right knee pooled).
Self-reported race was dichoto-mized (White or Caucasian/Other than
White) becausethe frequency in original groups ‘Black or African
Ameri-can’, ‘Asian’ and ‘other Non-White’ was low.Age, living
status, BMI, physical activity and comorbid
conditions were categorized because the distribution ofthe
covariates was skewed and in order to allow for non-linear
associations. Age was categorized into three groups(45–54, 55–64,
and 65–79 years). Education based on theOAI query refers to the
highest grade of school com-pleted, and had three categories:
primary/none (less thancollege), secondary (college graduate or
some graduateschool), and tertiary level (graduate degree).Living
status based on the OAI query was dichoto-
mized as living alone or living with someone else. BMIbased on
physical examination was categorized into fourgroups (< 25, 25
to < 30, 30 to < 35, ≥35 kg/m2). Physicalactivity was based
on the Physical Activity Scale for theElderly (PASE) score, where
higher scores indicate
greater physical activity [27], and was divided into quin-tiles
(0–90, 91–134, 135–175, 176–237, 238–526).Smoking status was
categorized into three groups(never/former/current smoker) based on
OAI questionson smoking pipe, cigars or cigarillos and smoking
ciga-rettes. Comorbid conditions based on Charlson comor-bidity
index score [28] was dichotomized (none/> 0comorbid
conditions).OAI data covered living status at baseline and year
3
follow-up visit, comorbid conditions at baseline, year 2and 4
follow-up visits, and smoking status at baselineand year 4
follow-up visit. Missing data regarding livingstatus, comorbid
conditions and smoking status wereimputed with the information
available from the baselinevisit or with the previous follow-up
visit (i.e., last obser-vation carried forward).
Statistical methodsDescriptive statistics were used to summarize
character-istics of OAI study participants. Repeated measures
gen-eralized estimating equation (GEE) models [29] wereused to
evaluate the population average [30] disutility inrelation to the
symptomatic and radiographic definitionsof knee OA in the
longitudinal data in order to take intoaccount the repeated
measurements of the same individ-ual as observations of a same
study subject are corre-lated. GEE modeling was used because it
allows theinclusion of all baseline participants in the analyses
ra-ther than only those participants who remained in thestudy and
had data available at all follow-up visits. Byutilizing a GEE
model, we were able to include data forparticipants who dropped out
of the study, or who didnot have radiographic knee examination in
all visits. AllGEE-models were specified using a normal
(Gaussian)distribution, an identity link function and an
unstructuredcorrelation matrix. The assumption of normal
distributionis justified by the large number of observations, and
be-cause by that means we avoided transformation of explana-tory
variable values equal to zero or one, which is requiredfor example
in gamma and beta regressions. The inspec-tion of disutility
distributions is presented in the additionalfile (see Additional
file 2). The repeated within-subjecttime-variable was the visit
number (0, 1, 2, 3 or 4). All GEEmodels were adjusted for the 11
covariates mentioned earl-ier. The dependent variable was the SF-6D
disutility scoreand the independent variable was a symptomatic or
radio-graphic definition of knee OA. Quasi Likelihood under
In-dependence Model Criterion was used to study models’goodness of
fit. Estimated marginal means and 95% Waldconfidence intervals of
disutility scores were compared be-tween different subgroups of
knee OA definitions. Partici-pants with no knee OA on the basis of
different definitionswere regarded as the reference group. Results
are pre-sented also as minimally important difference (MID)
values
Törmälehto et al. Health and Quality of Life Outcomes (2018)
16:154 Page 3 of 12
-
using a score difference of ≥0.027 as a cut-point for MID[31].
In all statistical analyses, conducted with IBM SPSSfor Windows,
version 23.0, the level of statistical signifi-cance was considered
as p < 0.05.
ResultsPopulation characteristicsThe flowchart of the cohort
definition and number of obser-vations eligible for the adjusted
GEE analyses is presented inFig. 1. In order to be eligible for
analyses, participants had tohave knee radiograph assessment
available for both kneesand to have answered the SF-12
questionnaire in full at thesame study visit at least once. In
addition, participants withpartial or total knee replacement (KR)
were excluded atbaseline. Altogether, 4278 participants were
eligible at base-line and had data on all adjusting variables.
Participants hav-ing a partial or total KR at least for one knee
during thefollow-up were censored once after the replacement.
Therewere 14,161 observations eligible for analyses accordingto the
radiographic definition of knee OA during the4-year follow-up. In
order to be eligible for analysesaccording to the symptomatic
definition of knee OA,participants had to have answered the knee
painquery also (14,074 observations eligible for analyses).The main
reason for missing data in the present ana-lyses was the missing
data on K-L grades for 42, 43and 44% of participants at study
visits 1, 2 and 3, re-spectively (see Fig. 1 and Additional file
3). Fifty-threepercent of participants eligible for the analyses
hadthree or more observations (see Additional file 3).The baseline
characteristics of participants are presented
in Table 1. At baseline, the mean age was 61.1 years(standard
deviation (SD) 9.2), and 58% of participantswere female. The mean
BMI was 28.5 (SD 4.8), and 76%of participants had no comorbid
conditions at baseline.Fifty-seven and 78% of participants reported
no knee in-juries and no knee surgical history, respectively, at
base-line. The characteristics of participants during follow-upare
presented in additional file (see Additional file 3).
Knee OA according to different definitionsThe prevalence of knee
OA according to different defini-tions among participants at
baseline and among all ob-servations are presented in Table 2. At
baseline, 57% ofparticipants had K-L grade ≥ 2 in at least one
knee, and26% had knee OA according to the symptomatic defin-ition
of knee OA. At baseline, when using the definitionof the highest
K-L grade for knee OA, the most preva-lent grades were 2 (30%) and
0 (28%), while the mostprevalent mean K-L grades were 0.0 (28%) and
2.0(17%). The most prevalent K-L grade combinations atbaseline were
(0;0) and (2;2) accounting for 28 and 13%of participants,
respectively. The prevalence of knee OA
according to different definitions during follow-up arepresented
in additional file (see Additional file 3).
SF-6D disutility scores in relation to definitions of
kneeosteoarthritisAt baseline, the mean (SD) unadjusted SF-6D
utilityscore of all study participants was 0.801 (0.120)
indicat-ing utility loss (disutility) equal to − 0.199 (see Table
1).The mean adjusted disutility scores in relation to differ-ent
symptomatic and radiological definitions of knee OAare presented in
Figs. 2, 3 and 4 and in more detail inadditional file (see
Additional file 4). The parameter esti-mates from GEE analyses are
presented in additional file(see Additional file 4). Pairwise
comparisons are pre-sented in additional file (see Additional file
5).The estimated disutility score increased with worsen-
ing of symptomatic or radiographic status of knee OA.Symptomatic
and radiographic knee OA in at least oneknee was associated with an
increase of 0.026 and 0.006(corresponding to 1.0 MID and 0.2 MID),
respectively,in disutility scores compared with the reference
group(Fig. 2a and b). Bilateral or unilateral symptomatic kneeOA
was associated with 0.030 (1.1 MID) or 0.024 (0.9MID) higher
disutility scores, respectively, comparedwith the reference group
(Fig. 2c). The estimated in-crease in disutility was 0.013 points
(0.5 MID) with bilat-eral radiographic knee OA compared with the
referencegroup (Fig. 2d).The estimated disutility scores increased
by 0.027 and
0.040 points (1.0 MID and 1.5 MID) from the best tothe worst
among the highest and mean of K-L grades(Fig. 3a and b),
respectively. Compared to a mean K-Lgrade value of 0.0, the
estimated increase in disutilityscore was statistically significant
with mean values of2.0, 3.0, 3.5, and 4.0 (Fig. 3b). An additional
file showsthis in more detail (see Additional file 5).Increase in
estimated disutility scores from the best
combination of K-L grades (0;0) to the worst (4;4) wasthe same
(0.040 points, corresponding to 1.5 MID) aswith mean K-L grade
classification due to arithmeticalreasons (Fig. 4). The disutility
scores in health states(2;2), (3;3), (4;0), (4;2), (4;3), and (4;4)
differed signifi-cantly from the reference health state (0;0) (Fig.
4). Anadditional file shows this in more detail (see Additionalfile
5).The difference in disutility scores between the worst
and best health state according to different definitions ofknee
OA was the greatest when using combinations ormean values of K-L
grades. The differences in theremaining definitions can be ranked
as follows: symp-tomatic OA (3-scale) > highest K-L grade >
symptomaticOA (2-scale) > radiographic OA (3-scale) >
radiographicOA (2-scale).
Törmälehto et al. Health and Quality of Life Outcomes (2018)
16:154 Page 4 of 12
-
DiscussionIn the present study, we estimated health disutilities
as-sociated with knee OA and its different symptomaticand
radiographic uni- and bilateral definitions. Our re-sults show that
the emergence of symptomatic knee OA
and increasing radiographic K-L grade are in relation
tostatistically and clinically significant worsening of
healthutility measured by the SF-6D instrument. The worsen-ing was
particulary strong when the subject had symp-tomatic bilateral knee
OA or a high mean value of K-L
Baseline
ParticipantsN=4796
Knee radiograph assessment N=4505& no knee replacement
N=4447
& answered the SF-12 questionnaire N=4395& data on all
adjusting variables N=4278
& answered the pain questionnaire N=4248*
For analysesN=4278
orN=4248*
Follow-up missed N=300
Year 1
ParticipantsN=4496
Knee radiograph assessment N=2596& no knee replacement
N=2537
& answered the SF-12 questionnaire N=2485& data on all
adjusting variables N=2424
& answered the pain questionnaire N=2409*(Comorbidities,
living and smoking statuses imputed)
For analysesN=2424
orN=2409*
Follow-up missed N=172
Year 2
ParticipantsN=4324
Knee radiograph assessment N=2447& no knee replacement
N=2359
& answered the SF-12 questionnaire N=2310& data on all
adjusting variables N=2249
& answered the pain questionnaire N=2230*(Living and smoking
statuses imputed)
For analysesN=2249
orN=2230*
Follow-up missed N=54
Year 3
ParticipantsN=4270
Knee radiograph assessment N=2354& no knee replacement
N=2215
& answered the SF-12 questionnaire N=2186& data on all
adjusting variables N=2059
& answered the pain questionnaire N=2053*(Comorbidities and
smoking status imputed)
For analysesN=2059
orN=2053*
Follow-up missed N=14
Year 4
ParticipantsN=4256
Knee radiograph assessment N=3610& no knee replacement
N=3431
& answered the SF-12 questionnaire N=3401& data on all
adjusting variables N=3151
& answered the pain questionnaire N=3134*(Living status
imputed)
For analysesN=3151
orN=3134*
Whole data
ObservationsN=22142
Knee radiograph assessment N=15512& no knee replacement
N=14989
& answered the SF-12 questionnaire N=14777& data on all
adjusting variables N=14161
& answered the pain questionnaire N=14074*
For analysesN=14161
orN=14074*
Fig. 1 The flow diagram of the OAI study participants and number
of observations. * Pain questionnaire required for the definition
ofsymptomatic knee OA (K-L grade≥ 2 and knee pain on more than half
the days during past month in the same knee)
Törmälehto et al. Health and Quality of Life Outcomes (2018)
16:154 Page 5 of 12
-
grades. Health disutility in bilateral knee OA differed
fromdisutility associated with unilateral knee OA when usingthe
radiographic definition. Based on our results, pain ex-perience is
an important factor in OA-related quality oflife as the disutility
score associated with knee OA wasgreater when taking pain
experience into account (i.e.,symptomatic definition) than when
using only radio-graphic definition (K-L grade ≥ 2).The results
indicate that symptomatic knee OA worsens
SF-6D health utility score significantly by 0.025 points
(0.9MID) on average. The association of unilateral and
bilateralsymptomatic knee OA with health utility were
similar.Previously, it has been shown that symptomatic knee OA
isassociated with a decrease of 0.003–0.19 points in EQ-5Dutility
scores [17, 18]. However, the precise comparabilityof the results
is challenging because of differentpreference-based health utility
measures [32] and knee OAdefinitions utilized in previous studies:
Muraki andco-authors [17] based their definition on the highest
K-Lgrade of both knees, similarly to one of our definitions.
Theyused K-L grade 3 as a cut-off point of radiographic kneeOA, and
symptomatic knee OA was defined as knee painlasting at least 1
month within the current or previous yearin the knee with K-L grade
≥ 3. Kiadaliri and co-authors [18]used clinical ACR (American
College of Rheumatology)criteria based on frequent knee pain,
crepitus, morningstiffness, age, and/or bony enlargements and a
radio-graphic definition that approximates K-L grade 2 orhigher.
Both of the studies utilized unilateral knee OA def-inition (i.e.,
the OA status of the other knee was omitted).We considered
unilateral symptomatic and radiographic
definitions (2-scale measures and the highest K-L grade)to be
overly broad definition of health states in a sense ofhealth
economics analyses. That is why we wanted toexamine, if we can
differentiate the changes in disutilityscores between bilateral
definitions and the five (0–4) K-Lgrades. Because health utility is
tied to a person’s generalexperience on quality of life but knee OA
may emergenceonly in one or both knees, we used bilateral
definitions ofOA and the mean and combination of K-L grades as
an
Table 1 Baseline population characteristics of
participantseligible for data analyses
Variable N (%) Mean (SD)
Participants 4278 (100.0)
Age, years 61.1 (9.2)
45–54 1264 (29.5)
55–64 1393 (32.6)
> 65 1621 (37.9)
Gender
Male 1782 (41.7)
Female 2496 (58.3)
Race
White or Caucasian 3458 (80.8)
Other 820 (19.2)
Educationa
Tertiary 1336 (31.2)
Secondary 1277 (29.9)
None / Primary 1665 (38.9)
Living status (number of persons) 1.3 (1.1)
Live alone 928 (21.7)
Living with someone else 3350 (78.3)
BMI, kg/m2 28.5 (4.8)
< 25 1040 (24.3)
25 to < 30 1690 (39.5)
30 to < 35 1122 (26.2)
≥35 426 (10.0)
Comorbidities 0.4 (0.8)
score 0 3254 (76.1)
score > 0 1024 (23.9)
Physical activity, PASEb 162.1 (81.8)
238–580 798 (18.7)
176–237 853 (19.9)
135–175 866 (20.2)
91–134 871 (20.4)
0–90 890 (20.8)
Smoking status
Never 1977 (46.2)
Former 1908 (44.6)
Current 393 (9.2)
Knee injuries
No 2424 (56.7)
Yes 1854 (43.3)
Table 1 Baseline population characteristics of
participantseligible for data analyses (Continued)
Variable N (%) Mean (SD)
Knee surgical history
No 3341 (78.1)
Yes 937 (21.9)
Disutility (SF-6D) score −0.199 (0.120)
Abbreviations: BMI body mass index; Comorbidites Charlson
comorbidity indexscore, PASE physical activity scale for the
elderly, SD standard deviationaThe highest grade of school
completed: tertiary (graduate degree), secondary(college graduate
or some graduate school), and primary/none level (lessthan
college)bPhysical activity PASE score quintiles (higher scores
indicate greaterphysical activity)
Törmälehto et al. Health and Quality of Life Outcomes (2018)
16:154 Page 6 of 12
-
attempt to take the effect of both knees on disutility
intoaccount, yet these are not validated and generally
acceptedoutcome measures. Our results indicate that as symptom-atic
definition was applied, the unilateral and bilateral defi-nitions
yielded similar outcomes. On the other hand,according to our
findings, we consider bilateral definitionimportant when plain
radiographic definition is applied.Our findings indicate that the
bilateral radiographic
definition of knee OA and the mean and
combinationclassifications of K-L grades were able to
differentiatedisutility scores between the health states although
painexperience was not included in the OA definition. MeanK-L grade
definition was able to differentiate moregroups than the highest
K-L grade. This may be result ofthe fact that K-L grades of both
knees were included,even though rather heterogeneous K-L pairs
result in thesame mean value. The advantage of combination
K-Lgrades is that it does neither omit nor summarize theK-L grades
of the knees, but it resulted in smaller groupsizes, which enlarged
the reported confidence intervals.The capability of K-L grade alone
to differentiate the
disutility between sequential health states was limited,although
the difference between the extreme K-L gradeswas explicit in all
definitions. For example, the definitionbased on the highest K-L
grade was able to differentiateall pairwise disutility comparisons
only with K-L grade 4.The reason for inadequate differentiation of
disutility be-tween the consecutive K-L grades may be that the
radio-graphic classification has its shortcomings, and K-L ismore
likely an ordinal, not an interval scale. Firstly, theoriginal K-L
grades are verbally described, and the inter-pretation of them and
radiographs is inconsistent andvaries between observers [33].
Secondly, radiographicfindings of knee OA does not absolutely
result in painexperience or disability [34], which are the obvious
fac-tors deteriorating the patient-reported HRQoL. On theother
hand, radiographic findings have some prognosticstrength in
relation to the symptomatology [35, 36].Results of the present
study confirm that the bilateral
definition of knee OA was related to higher disutilityscores if
the definition was symptomatic, incorporatingpatient-reported
frequent pain experience, rather thansimple radiographic definition
[17, 18]. Interestingly, ourresults are also consistent with
previous findings that
Table 2 Prevalence of knee OA according to
differentdefinitions
Variable Baseline N (%) All observations N (%)
N 4278 14,161
Symptomatic OA statusa (2-scale)
No 3134 (73.8) 9538 (67.8)
Yes, uni- or bilateral 1114 (26.2) 4536 (32.2)
Symptomatic OA statusa (3-scale)
No 3134 (73.8) 9538 (67.8)
Yes, unilateral 762 (17.9) 2999 (21.3)
Yes, bilateral 352 (8.3) 1537 (10.9)
K-L grade≥ 2 (2-scale)
No 1860 (43.5) 3395 (24.0)
Yes, uni- or bilateral 2418 (56.5) 10,766 (76.0)
K-L grade≥ 2 (3-scale)
No 1860 (43.5) 3395 (24.0)
Yes, unilateral 1112 (26.0) 4785 (33.8)
Yes, bilateral 1306 (30.5) 5981 (42.2)
The highest K-L grade
0 1199 (28.0) 2051 (14.5)
1 661 (15.5) 1344 (9.5)
2 1288 (30.1) 5544 (39.1)
3 841 (19.7) 3693 (26.1)
4 289 (6.8) 1529 (10.8)
Mean of K-L grades
0.0 1199 (28.0) 2051 (14.5)
0.5 383 (9.0) 774 (5.5)
1.0 612 (14.3) 2032 (14.3)
1.5 550 (12.9) 2346 (16.6)
2.0 728 (17.0) 3151 (22.3)
2.5 391 (9.1) 1660 (11.7)
3.0 313 (7.3) 1470 (10.4)
3.5 100 (2.3) 566 (4.0)
4.0 2 (0.0) 111 (0.8)
Combination of K-L grades
(0;0) 1199 (28.0) 2051 (14.5)
(1;0) 383 (9.0) 774 (5.5)
(1;1) 278 (6.5) 570 (4.0)
(2;0) 334 (7.8) 1462 (10.3)
(2;1) 403 (9.4) 1714 (12.1)
(2;2) 551 (12.9) 2368 (16.7)
(3;0) 147 (3.4) 632 (4.5)
(3;1) 123 (2.9) 525 (3.7)
(3;2) 340 (7.9) 1466 (10.4)
(3;3) 231 (5.4) 1070 (7.6)
(4;0) 54 (1.3) 258 (1.8)
Table 2 Prevalence of knee OA according to differentdefinitions
(Continued)
Variable Baseline N (%) All observations N (%)
(4;1) 51 (1.2) 194 (1.4)
(4;2) 82 (1.9) 400 (2.8)
(4;3) 100 (2.3) 566 (4.0)
(4;4) 2 (0.0) 111 (0.8)aK-L grade ≥ 2 and knee pain on more than
half the days during past month inthe same knee
Törmälehto et al. Health and Quality of Life Outcomes (2018)
16:154 Page 7 of 12
-
Fig. 2 SF-6D disutility scores in relation to symptomatic and
radiographic knee OA definitions. a 2-scale symptomatic knee OA
(K-L grade≥ 2 andfrequent knee pain in the same knee in at least
one knee) b 2-scale radiographic knee OA (K-L grade≥ 2 in at least
one knee) c 3-scale symptomaticknee OA (K-L grade≥ 2 and frequent
knee pain in the same knee) d 3-scale radiographic knee OA (K-L
grade≥ 2). Disutility of the best health statepossible (no OA) is
set as the horizontal reference line (dotted). Other horizontal
lines (dotted) are set in 0.027 point intervals representing 1.0
MID(minimally important difference). Error bars are 95% confidence
intervals. The mean difference is significant at the 0.05 level (*)
in comparison to thebest health state (no OA)
Fig. 3 SF-6D disutility scores in relation to 5- and 9-scale
radiographic knee OA definitions. a 5-scale radiographic knee OA
(the highest K-L grade inboth knees) or b 9-scale radiographic knee
OA (mean of K-L grades in both knees). Disutility of the best
health state possible (K-L grade 0) is set as thehorizontal
reference line (dotted). Other horizontal lines (dotted) are set in
0.027 point intervals representing 1.0 MID (minimally important
difference).Error bars are 95% confidence intervals. The mean
difference is significant at the 0.05 level (*) in comparison to
the best health state (K-L grade 0)
Törmälehto et al. Health and Quality of Life Outcomes (2018)
16:154 Page 8 of 12
-
there is a relationship between self-reported knee painand
radiological OA defined by the highest K-L grade ofboth knees based
on a previous study showing that kneepain experience differed by
K-L grades and was explicitat least with the worst K-L grade
[35].Results of the present study may also be interpreted in
the perspective of a health economic evaluation. If
anintervention would prevent the progression of jointdamage in both
knees from the mean K-L grade 0 to 2or from 2 to 4 the raw QALY
gain would be 0.012, or0.028 per year, respectively. In the
treatment of OA,long-term objectives should focus on prevention of
jointdamage and on the improvement of quality of life [37].If there
was a more consistent diagnostic test for knee
OA, defining the health states would be more straight-forward.
At present, the comparability of clinical andcost-effectiveness
studies is challenging due to differentdefinitions of knee OA. The
comparability between studiesis also challenging because of
different applied health utilityinstruments. The present study
applied the SF-12 question-naire, and the results were converted to
SF-6D-utilities by amapping algorithm [20]. SF-6D takes into
account six di-mensions of the original SF-12 questionnaire. An
advantageof the SF-6D measure is that it provides a single index
scorefor the estimation of QALYs in cost-effectiveness analyses.We
used the value of 0.027 [31] as a cut-point for MID as asmallest
change in health utility that is important to a sub-ject. However,
estimates for SF-6D MID-values vary from0.010 to 0.048 [38].To our
knowledge, the present study is the first to quan-
tify the health disutility related to knee OA using the
gen-eric, preference-based SF-6D instrument in relation todifferent
symptomatic and radiographic uni- and bilateraldefinitions of knee
OA. Although the impact of knee OAon quality of life has been
studied widely, there has beenlack of information on
preference-based quality of life inrelation to different
definitions of knee OA, and particu-larly in relation to uni- and
bilateral definitions. Previously,
it has been shown that bilateral knee pain is associatedwith
poorer quality of life [10]. However, we found that bi-lateral knee
OA definition is uncommon in HRQoL re-search, even though both
knees naturally affect patients’quality of life. This is probably a
result of the fact that thegenerally used definitions of knee OA do
not differentiatebetween uni- or bilateral knee OA. In the previous
kneeOA related quality of life, functional decline, and pain
stud-ies, the use of unilateral knee OA definition is usually
im-plicitly expressed, and the selection of the analyzed kneehas
been variable (e.g., the worse knee or symptomatic def-inition with
knee OA in at least one knee or knee ran-domly selected or both
knees analyzed separately) (e.g.,[17, 18, 25, 35, 39]). However, in
our study we explicitlyquantified the health disutility in relation
to symptomaticand radiographic uni- and bilateral knee OA
definitions.The strength of this study is that we quantified
the
SF-6D scores in a repeated measures design with a sub-stantial
number of OAI study subjects. OAI is a largeopen access database
with relatively low rate ofdrop-outs and high response rate to the
validated SF-12questionnaire. Earlier studies have quantified
health util-ity in relation to knee OA using EQ-5D instrument
incross-sectional study design [17, 18]. There is no goldstandard
for generic preference-based HRQoL measure.National health
technology assessment organizationssuch as NICE (National Institute
for Health and CareExcellence) recommend EQ-5D as a generic health
util-ity measure. However, a working group for standard setof
outcome measures recently recommended both SF-12and EQ-5D as
standard patient reported outcome meas-ure tools for persons with
hip or knee OA for HRQoLevaluation [24]. Variety of
disease-spesific instruments(KOOS, WOMAC, knee pain) may also be
used tomeasure knee OA related quality of life. The OAI data-set
includes KOOS, WOMAC and SF-12 questionnaires.Although the
disease-spesific instruments have advan-tages, and they cover the
dimensions relevant to knee OA,
Fig. 4 SF-6D disutility scores in relation to 15-scale
radiographic knee OA definition. 15-scale radiographic knee OA
(combination of K-L grades in bothknees). Disutility of the best
health state possible (K-L grade 0;0) is set as the horizontal
reference line (dotted). Other horizontal lines (dotted) are set
in0.027 point intervals representing 1.0 MID (minimally important
difference). Error bars are 95% confidence intervals. The mean
difference is significantat the 0.05 level (*) in comparison to the
best health state (K-L grade 0;0)
Törmälehto et al. Health and Quality of Life Outcomes (2018)
16:154 Page 9 of 12
-
they are not suitable for calculating QALYs for health eco-nomic
analyses unlike generic, preference-based healthutility
instruments. However, there are mapping algo-rithms available to
estimate the preference-based utilityfrom the disease-spesific
instrument scores but theyshould be applied cautiously [40].The
results of the present study should be interpreted in
light of some limitations. Firstly, while we have reportedthe
health disutilities of OAI study participants, the findingsof this
study are, however, based on UK population-basedpreferences. There
is evidence that health state valuing maydiffer between countries
[41, 42]. SF-6D valuation surveyshave been completed also in China
[43], Japan [41],Portugal [44], Spain [45] and Brazil [46].
Secondly, as wetook patient reported comorbid conditions into
account wedid not specifically focus on comorbidities of
musculoskel-etal system as adjusting variables. However, problems
inmusculoskeletal system, joint pain comorbidity, coexistingback
pain, and hip OA have been reported to have impacton the quality of
life and experience of knee pain in patientswith knee OA [8,
47–50]. Thirdly, we used symptomaticdefinition only in 2- and
3-scale definitions. It would havebeen interesting to combine pain
status also with the differ-ent radiographic definitions, but this
would have resulted intoo small group sizes. Fourthly, there was a
substantialnumber of missing K-L grades at year 1, 2 and 3
follow-upvisits. The GEE model utilized has a straightforward
as-sumption of missing completely at random (MCAR). Themissing K-L
grades could have been ‘safely’ imputed withthe technique of ‘last
observation carried forward’ as thedegeneration in knee joint (K-L
grade) is not assumed torecover. However, 94 and 85% of
participants had K-Lgrade available at baseline and at year 4
follow-up visit, re-spectively. Fifthly, the OAI study participants
may behealthier and more educated than general population,which may
limit the generalizability of the findings.
ConclusionsThe results of the present study confirm that knee OA
isassociated with diminished HRQoL. Our study expandsupon the
previous research by quantifying the health dis-utility values
associated with different symptomatic andradiographic definitions
of knee OA, as the gold standardof knee OA definition is currently
unavailable. As previousresearch, we found different definitions of
knee OA to beassociated with different health disutilities, which
can havean effect on the results of health economic analyses.
Healthdisutility in relation to bilateral knee OA is less studied,
andwe found it to differ from disutility associated with
unilat-eral knee OA when using radiographic definition of kneeOA.
The radiographic definition was, however, a crudemeasure to
differentiate the disutility by sequential K-Lgrades. The
performance of symptomatic definition was
better, indicating that pain experience is an important fac-tor
in knee OA related quality of life.
Additional files
Additional file 1 OAI datasets. This table reports specific OAI
datasets,from where data for each subject were collected and,
applied in analyses.(PDF 10 kb)
Additional file 2 Histograms of SF-6D score frequencies. These
figuresreport distribution of SF-6D disutility scores presented as
histograms andnormal curves. (PDF 682 kb)
Additional file 3 Characteristics of participants and prevalence
of kneeOA. These tables report characteristics of participants
eligible for dataanalyses and prevalence of knee OA according to
different definitionsduring follow-up. (PDF 187 kb)
Additional file 4 SF-6D-disutility scores and parameter
estimates. Thesetables reports estimated marginal means of
SF-6D-disutility scores and95% confidence intervals (CI) from
GEE-analyses. (PDF 294 kb)
Additional file 5 Pairwise differences between SF-6D-disutility
scores.These tables report pairwise differences between estimated
marginalmeans of SF-6D-disutility scores from GEE-analyses. (PDF
151 kb)
AbbreviationsACR: American College of Rheumatology; BMI: Body
mass index; EQ-5D: EuroQol-5 dimensions questionnaire; GEE:
Generalized estimatingequation model; HRQoL: Health-related quality
of life; K-L: Kellgren-Lawrenceradiographic system; KOOS: Knee
injury and osteoarthritis outcome score;MCAR: Missing completely at
random; MID: Minimally important difference;NICE: National
Institute for Health and Care Excellence; OA: Osteoarthritis;OAI:
Osteoarthritis Initiative; PASE: Physical activity scale for the
elderly score;QALY: Quality-adjusted life year; SD: Standard
deviation; SF-12: Short formhealth survey; WOMAC: Western Ontario
and McMaster Universitiesosteoarthritis index
AcknowledgementsThe authors wish to thank the patients and staff
of all the hospitals who havecontributed data to the Osteoarthritis
Initiative. The OAI is a public-privatepartnership comprised of
five contracts (N01-AR-2-2258; N01-AR-2-2259; N01-AR-2-2260;
N01-AR-2-2261; N01-AR-2-2262) funded by the National Institutes
ofHealth, a branch of the Department of Health and Human Services,
andconducted by the OAI Study Investigators. Private funding
partners includeMerck Research Laboratories; Novartis
Pharmaceuticals Corporation,GlaxoSmithKline; and Pfizer, Inc.
Private sector funding for the OAI is managedby the Foundation for
the National Institutes of Health. This manuscript wasprepared
using an OAI public use data set and does not reflect the opinions
orviews of the OAI investigators, the NIH, or the private funding
partners. Theauthors wish to thank biostatistician Tuomas Selander,
MSc, and statistician PiiaLavikainen, PhD, for their guidance in
statistical analyses.
FundingThe research leading to these results has received
funding from theAcademy of Finland (grant 305138) and the European
Research Council(ERC) under the European Union’s Horizon 2020
research and innovationprogramme (grant agreement No 755037). The
institution had no role in thedesign of the study or collection,
analysis and interpretation of data, or inwriting the manuscript,
or in the decision to submit the manuscript forpublication.
Availability of data and materialsThe data that support the
findings of this study are available from theOsteoarthritis
Initiative (OAI), http://www.oai.ucsf.edu/.
Authors’ contributionsAll authors were involved in conception
and designing the study. STundertook the analyses, interpretation
of the data, and drafting themanuscript. All authors contributed to
revising the manuscript critically forimportant intellectual
content, and have approved the final version to besubmitted.
Törmälehto et al. Health and Quality of Life Outcomes (2018)
16:154 Page 10 of 12
https://doi.org/10.1186/s12955-018-0979-7https://doi.org/10.1186/s12955-018-0979-7https://doi.org/10.1186/s12955-018-0979-7https://doi.org/10.1186/s12955-018-0979-7https://doi.org/10.1186/s12955-018-0979-7http://www.oai.ucsf.edu/
-
Ethics approval and consent to participateThis article was
prepared using an Osteoarthritis Initiative (OAI) public-usedata
set (http://www.oai.ucsf.edu/). Ethical approval for collecting all
subjectinformation was provided by the OAI, and informed consent
was obtainedfrom all individual participants included in the study.
This article does notcontain any studies with human participants
performed by any of the authors.Access, download and analyses of
the OAI data were performed under the OAIData Use Agreement.
Consent for publicationNot applicable
Competing interestsST has been a paid employee of Medfiles Ltd.,
and is the founder ofMediSoili Oy. JM is a founding partner of
ESiOR Oy. These companies werenot involved in carrying out this
research. All other authors declare that theyhave no competing
interests.
Publisher’s NoteSpringer Nature remains neutral with regard to
jurisdictional claims inpublished maps and institutional
affiliations.
Author details1Pharmacoeconomics and Outcomes Research Unit
(PHORU), School ofPharmacy, University of Eastern Finland, Kuopio,
Finland. 2Department ofApplied Physics, University of Eastern
Finland, Kuopio, Finland. 3Institute ofBiomedicine, University of
Turku, Turku, Finland. 4Department of Physical andRehabilitation
Medicine, Helsinki University Hospital, Helsinki,
Finland.5University of Helsinki, Helsinki, Finland. 6Diagnostic
Imaging Centre, KuopioUniversity Hospital, Kuopio, Finland.
Received: 31 January 2018 Accepted: 18 July 2018
References1. Lawrence RC, Felson DT, Helmick CG, Arnold LM, Choi
H, Deyo RA, et al.
National Arthritis Data Workgroup. Estimates of the prevalence
of arthritisand other rheumatic conditions in the United States.
Part II. Arthritis Rheum.2008;
https://doi.org/10.1002/art.23176.
2. GBD 2015 Disease and Injury Incidence and Prevalence
Collaborators.Global, regional, and national incidence, prevalence,
and years lived withdisability for 310 diseases and injuries,
1990–2015: a systematic analysis forthe Global Burden of Disease
Study 2015. Lancet:2016.
https://doi.org/10.1016/S0140-6736(16)31678-6.
3. Xie F, Kovic B, Jin X, He X, Wang M, Silvestre C. Economic
and humanisticburden of osteoarthritis: a systematic review of
large sample studies.Pharmacoeconomics. 2016;
https://doi.org/10.1007/s40273-016-0424-x.
4. Zheng H, Chen C. Body mass index and risk of knee
osteoarthritis:systematic review and meta-analysis of prospective
studies. BMJ Open.2015;
https://doi.org/10.1136/bmjopen-2014-007568.
5. Blagojevic M, Jinks C, Jeffery A, Jordan KP. Risk factors for
onset ofosteoarthritis of the knee in older adults: a systematic
review and meta-analysis. Osteoarthr Cartil. 2010;
https://doi.org/10.1016/j.joca.2009.08.010.
6. Ackerman IN, Bucknill A, Page RS, Broughton NS, Roberts C,
Cavka B, et al.The substantial personal burden experienced by
younger people with hipor knee osteoarthritis. Osteoarthr Cartil.
2015; https://doi.org/10.1016/j.joca.2015.04.008.
7. Saarni SI, Härkänen T, Sintonen H, Suvisaari J, Koskinen S,
Aromaa A, et al.The impact of 29 chronic conditions on
health-related quality of life: ageneral population survey in
Finland using 15D and EQ-5D. Qual Life Res.2006;
https://doi.org/10.1007/s11136-006-0020-1.
8. Hoogeboom TJ, den Broeder AA, de Bie RA, van den Ende CH.
Longitudinalimpact of joint pain comorbidity on quality of life and
activity levels in kneeosteoarthritis: data from the osteoarthritis
initiative. Rheumatology (Oxford).2013;
https://doi.org/10.1093/rheumatology/kes314.
9. Conaghan PG, Peloso PM, Everett SV, Rajagopalan S, Black CM,
Mavros P,et al. Inadequate pain relief and large functional loss
among patients withknee osteoarthritis: evidence from a prospective
multinational longitudinalstudy of osteoarthritis real-world
therapies. Rheumatology (Oxford).
2015;https://doi.org/10.1093/rheumatology/keu332.
10. Bindawas SM, Vennu V, Al SS. Differences in health-related
quality of lifeamong subjects with frequent bilateral or unilateral
knee pain: data fromthe osteoarthritis initiative study. J Orthop
Sports Phys Ther. 2015;
https://doi.org/10.2519/jospt.2015.5123.
11. Shan L, Shan B, Suzuki A, Nouh F, Saxena A. Intermediate and
long-termquality of life after total knee replacement: a systematic
review and meta-analysis. J Bone Joint Surg Am. 2015;
https://doi.org/10.2106/JBJS.M.00372.
12. Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW.
Validation studyof WOMAC: a health status instrument for measuring
clinically importantpatient relevant outcomes to antirheumatic drug
therapy in patients withosteoarthritis of the hip or knee. J
Rheumatol. 1988;15:1833–40.
13. Roos EM, Roos HP, Lohmander LS, Ekdahl C, Beynnon BD. Knee
injury andOsteoarthritis Outcome Score (KOOS)--development of a
self-administeredoutcome measure. J Orthop Sports Phys Ther. 1998;
https://doi.org/10.2519/jospt.1998.28.2.88.
14. Drummond MF, Sculpher MJ, Torrance GW, O’Brien BJ, Stoddart
GL. Cost-effectiveness analysis. In: Methods for the economic
evaluation of healthcare Programmes. New York: Oxford University
Press Inc; 2005. p. 103–36.
15. Torrance GW, Feeny D. Utilities and quality-adjusted life
years. Int J TechnolAssess Health Care. 1989;5:559–75.
16. Schiphof D, de Klerk BM, Koes BW, Bierma-Zeinstra S. Good
reliability,questionable validity of 25 different classification
criteria of kneeosteoarthritis: a systematic appraisal. J Clin
Epidemiol. 2008;
https://doi.org/10.1016/j.jclinepi.2008.04.003.
17. Muraki S, Akune T, Oka H, En-yo Y, Yoshida M, Saika A, et
al. Association ofradiographic and symptomatic knee osteoarthritis
with health-relatedquality of life in a population-based cohort
study in Japan: the ROAD study.Osteoarthr Cartil. 2010;
https://doi.org/10.1016/j.joca.2010.06.001.
18. Kiadaliri AA, Lamm CJ, de Verdier MG, Engström G, Turkiewicz
A,Lohmander LS, et al. Association of knee pain and different
definitions ofknee osteoarthritis with health-related quality of
life: a population-basedcohort study in southern Sweden. Health
Qual Life Outcomes. 2016;
https://doi.org/10.1186/s12955-016-0525-4.
19. Nevitt MC, Felson DT, Lester G. The Osteoarthritis
Initiative. In: Protocol forthe cohort study. National Institute of
Arthritis, Musculoskeletal and SkinDiseases. V 1.1 6.21.06.
https://oai.epi-ucsf.org/datarelease/About.asp.Accessed 26 Nov
2016.
20. Brazier JE, Roberts J. The estimation of a preference-based
measure ofhealth from the SF-12. Med Care. 2004;42:851–9.
21. Kellgren JH, Lawrence JS. Radiological assessment of
osteo-arthrosis. AnnRheum Dis. 1957;16:494–502.
22. Jordan JM, Helmick CG, Renner JB, Luta G, Dragomir AD,
Woodard J, et al.Prevalence of knee symptoms and radiographic and
symptomatic kneeosteoarthritis in African Americans and Caucasians:
the Johnston Countyosteoarthritis project. J Rheumatol.
2007;34:172–80.
23. Turkiewicz A, Gerhardsson de Verdier M, Engström G, Nilsson
PM, MellströmC, Lohmander LS, et al. Prevalence of knee pain and
knee OA in southernSweden and the proportion that seeks medical
care. Rheumatology(Oxford). 2015;
https://doi.org/10.1093/rheumatology/keu409.
24. Rolfson O, Wissig S, van Maasakkers L, Stowell C, Ackerman
I, Ayers D, et al.Defining an international standard set of outcome
measures for patientswith hip or knee osteoarthritis: consensus of
the international consortiumfor health outcomes measurement hip and
knee osteoarthritis workinggroup. Arthritis Care Res (Hoboken).
2016; https://doi.org/10.1002/acr.22868.
25. Collins JE, Katz JN, Dervan EE, Losina E. Trajectories and
risk profiles of painin persons with radiographic, symptomatic knee
osteoarthritis: data fromthe osteoarthritis initiative. Osteoarthr
Cartil. 2014; https://doi.org/10.1016/j.joca.2014.03.009.
26. Reichmann WM, Katz JN, Kessler CL, Jordan JM, Losina E.
Determinantsof self-reported health status in a population-based
sample of personswith radiographic knee osteoarthritis. Arthritis
Rheum. 2009; https://doi.org/10.1002/art.24839.
27. Washburn RA, Smith KW, Jette AM, Janney CA. The physical
activityscale for the elderly (PASE): development and evaluation. J
ClinEpidemiol. 1993;46:153–62.
28. Katz JN, Chang LC, Sangha O, Fossel AH, Bates DW. Can
comorbidity bemeasured by questionnaire rather than medical record
review? Med Care.1996;34:73–84.
29. Hanley JA, Negassa A, Edwardes MD, Forrester JE. Statistical
analysis ofcorrelated data using generalized estimating equations:
an orientation. AmJ Epidemiol. 2003;157:364–75.
Törmälehto et al. Health and Quality of Life Outcomes (2018)
16:154 Page 11 of 12
http://www.oai.ucsf.edu/https://doi.org/10.1002/art.23176https://doi.org/10.1016/S0140-6736(16)31678-6https://doi.org/10.1016/S0140-6736(16)31678-6https://doi.org/10.1007/s40273-016-0424-xhttps://doi.org/10.1136/bmjopen-2014-007568https://doi.org/10.1016/j.joca.2009.08.010https://doi.org/10.1016/j.joca.2015.04.008https://doi.org/10.1016/j.joca.2015.04.008https://doi.org/10.1007/s11136-006-0020-1https://doi.org/10.1093/rheumatology/kes314https://doi.org/10.1093/rheumatology/keu332https://doi.org/10.2519/jospt.2015.5123https://doi.org/10.2519/jospt.2015.5123https://doi.org/10.2106/JBJS.M.00372https://doi.org/10.2519/jospt.1998.28.2.88https://doi.org/10.2519/jospt.1998.28.2.88https://doi.org/10.1016/j.jclinepi.2008.04.003https://doi.org/10.1016/j.jclinepi.2008.04.003https://doi.org/10.1016/j.joca.2010.06.001https://doi.org/10.1186/s12955-016-0525-4https://doi.org/10.1186/s12955-016-0525-4https://oai.epi-ucsf.org/datarelease/About.asphttps://doi.org/10.1093/rheumatology/keu409https://doi.org/10.1002/acr.22868https://doi.org/10.1016/j.joca.2014.03.009https://doi.org/10.1016/j.joca.2014.03.009https://doi.org/10.1002/art.24839https://doi.org/10.1002/art.24839
-
30. Twisk JW. Longitudinal data analysis. A comparison between
generalizedestimating equations and random coefficient analysis.
Eur J Epidemiol.2004;19:769–76.
31. Luo N, Johnson J, Coons SJ. Using instrument-defined health
state transitionsto estimate minimally important differences for
four preference-based health-related quality of life instruments.
Med Care. 2010;48:365–71.
32. Brazier J, Roberts J, Tsuchiya A, Busschbach J. A comparison
of theEQ-5D and SF-6D across seven patient groups. Health Econ.
2004;https://doi.org/10.1002/hec.866.
33. Spector TD, Cooper C. Radiographic assessment of
osteoarthritis in populationstudies: whither Kellgren and Lawrence?
Osteoarthr Cartil. 1993;1:203–6.
34. Bedson J, Croft PR. The discordance between clinical and
radiographic kneeosteoarthritis: a systematic search and summary of
the literature. BMCMusculoskelet Disord. 2008;
https://doi.org/10.1186/1471-2474-9-116.
35. Laxafoss E, Jacobsen S, Gosvig KK, Sonne-Holm S. Case
definitions of kneeosteoarthritis in 4,151 unselected subjects:
relevance for epidemiologicalstudies: the Copenhagen osteoarthritis
study. Skelet Radiol. 2010;
https://doi.org/10.1007/s00256-009-0856-x.
36. Pereira D, Severo M, Santos RA, Barros H, Branco J, Lucas R,
et al. Knee and hipradiographic osteoarthritis features:
differences on pain, function and quality oflife. Clin Rheumatol.
2016; https://doi.org/10.1007/s10067-015-3087-7.
37. Bijlsma JW, Berenbaum F, Lafeber FP. Osteoarthritis: an
update withrelevance for clinical practice. Lancet. 2011;
https://doi.org/10.1016/S0140-6736(11)60243-2.
38. Walters SJ, Brazier JE. Comparison of the minimally
important difference fortwo health state utility measures: EQ-5D
and SF-6D. Qual Life Res.2005;14:1523–32.
39. White DK, Neogi T, Nguyen US, Niu J, Zhang Y. Trajectories
of functionaldecline in knee osteoarthritis: the osteoarthritis
initiative. Rheumatology(Oxford). 2016;
https://doi.org/10.1093/rheumatology/kev419.
40. Kiadaliri AA, Englund M. Assessing the external validity of
algorithms toestimate EQ-5D-3L from the WOMAC. Health Qual Life
Outcomes. 2016;https://doi.org/10.1186/s12955-016-0547-y.
41. Tsuchiya A, Ikeda S, Ikegami N, Nishimura S, Sakai I, Fukuda
T, et al.Estimating an EQ-5D population value set: the case of
Japan. Health Econ.2002; https://doi.org/10.1002/hec.673.
42. Brazier JE, Fukuhara S, Roberts J, Kharroubi S, Yamamoto Y,
Ikeda S, et al.Estimating a preference-based index from the
Japanese SF-36. J ClinEpidemiol. 2009;
https://doi.org/10.1016/j.jclinepi.2009.01.022.
43. Lam CL, Brazier J, McGhee SM. Valuation of the SF-6D health
states isfeasible, acceptable, reliable, and valid in a Chinese
population. ValueHealth. 2008;
https://doi.org/10.1111/j.1524-4733.2007.00233.x.
44. Ferreira LN, Ferreira PL, Pereira LN, Brazier J, Rowen D. A
Portuguese valueset for the SF-6D. Value Health. 2010;
https://doi.org/10.1111/j.1524-4733.2010.00701.x.
45. Méndez I, Abellán Perpiñán JM, Sánchez Martínez FI, Martínez
Pérez JE.Inverse probability weighted estimation of social tariffs:
an illustration usingthe SF-6D value sets. J Health Econ. 2011;
https://doi.org/10.1016/j.jhealeco.2011.07.013.
46. Cruz LN, Camey SA, Hoffmann JF, Rowen D, Brazier JE, Fleck
MP, et al. Estimatingthe SF-6D value set for a population-based
sample of Brazilians. Value Health.2011;14(Suppl 1):S108–14.
https://doi.org/10.1016/j.jval.2011.05.012.
47. Bindawas SM, Vennu V, Auais M. Health-related quality of
life in older adultswith bilateral knee pain and back pain: data
from the osteoarthritis initiative.Rheumatol Int. 2015;
https://doi.org/10.1007/s00296-015-3309-y.
48. Wesseling J, Welsing PM, Bierma-Zeinstra SM, Dekker J,
Gorter KJ,Kloppenburg M, et al. Impact of self-reported comorbidity
on physical andmental health status in early symptomatic
osteoarthritis: the CHECK (CohortHip and Cohort Knee) study.
Rheumatology (Oxford). 2013;
https://doi.org/10.1093/rheumatology/kes288.
49. Muraki S, Akune T, Oka H, En-Yo Y, Yoshida M, Saika A, et
al. Health-relatedquality of life in subjects with low back pain
and knee pain in a population-based cohort study of Japanese men:
the Research on OsteoarthritisAgainst Disability study. Spine
(Phila Pa 1976). 2011;
https://doi.org/10.1097/BRS.0b013e3181fa60d1.
50. Schiphof D, Kerkhof HJ, Damen J, de Klerk BM, Hofman A, Koes
BW, et al.Factors for pain in patients with different grades of
knee osteoarthritis.Arthritis Care Res (Hoboken). 2013;
https://doi.org/10.1002/acr.21886.
Törmälehto et al. Health and Quality of Life Outcomes (2018)
16:154 Page 12 of 12
https://doi.org/10.1002/hec.866https://doi.org/10.1186/1471-2474-9-116https://doi.org/10.1007/s00256-009-0856-xhttps://doi.org/10.1007/s00256-009-0856-xhttps://doi.org/10.1007/s10067-015-3087-7https://doi.org/10.1016/S0140-6736(11)60243-2https://doi.org/10.1016/S0140-6736(11)60243-2https://doi.org/10.1093/rheumatology/kev419https://doi.org/10.1186/s12955-016-0547-yhttps://doi.org/10.1002/hec.673https://doi.org/10.1016/j.jclinepi.2009.01.022https://doi.org/10.1111/j.1524-4733.2007.00233.xhttps://doi.org/10.1111/j.1524-4733.2010.00701.xhttps://doi.org/10.1111/j.1524-4733.2010.00701.xhttps://doi.org/10.1016/j.jhealeco.2011.07.013https://doi.org/10.1016/j.jhealeco.2011.07.013https://doi.org/10.1016/j.jval.2011.05.012https://doi.org/10.1007/s00296-015-3309-yhttps://doi.org/10.1093/rheumatology/kes288https://doi.org/10.1093/rheumatology/kes288https://doi.org/10.1097/BRS.0b013e3181fa60d1https://doi.org/10.1097/BRS.0b013e3181fa60d1https://doi.org/10.1002/acr.21886
AbstractBackgroundMethodsResultsConclusions
BackgroundMethodsData sourcesPreference-based health utility
indexDefinitions of knee OACovariatesStatistical methods
ResultsPopulation characteristicsKnee OA according to different
definitionsSF-6D disutility scores in relation to definitions of
knee osteoarthritis
DiscussionConclusionsAdditional
filesAbbreviationsAcknowledgementsFundingAvailability of data and
materialsAuthors’ contributionsEthics approval and consent to
participateConsent for publicationCompeting interestsPublisher’s
NoteAuthor detailsReferences